Tab B goes out. The numbers are good.

Algorithms train taste through dopamine. Every time the optimized version performs better, your editing brain learns 'that direction = correct.' Repeated over months, the gap between Tab A and Tab B closes — not because you got better at writing, but because you stopped being able to write Tab A. The drift is invisible while it's happening. The 2026 question isn't 'is using AI cheating?' It's 'is the algorithm doing the editing inside my head?' Tonight: numbers go up. A small flat feeling you can't name. In a month: harder to write the Tab A version even when you try. In a year: your audience is bigger and the work is more like everyone else's.

Tab A goes out. The numbers are smaller. The work is yours.

Voice isn't a fixed thing you have or don't have. It's a muscle. Each piece you publish in your actual voice trains it; each piece you publish in the algorithm's voice un-trains it. The 'authentic' creators a few years from now will be the ones who refused the optimization curve enough times that their voice didn't drift. That's the actual moat in a 2026 content landscape where competence is cheap and voice is rare. Tonight: smaller numbers, slightly more uncomfortable, more on your side. In a month: the Tab A version comes more easily next time. In a year: a body of work that feels coherent and unmistakable.

You make a third version: Tab A's voice, Tab B's clarity edits.

The hybrid is the most defensible *if* you're honest about which moves came from where. The failure mode is calling Tab B 'a hybrid' because you kept three of your sentences. The cleanest test: would you publish this draft if no one was going to see the metrics? If yes, it's yours. If you hesitate, it's still mostly Tab B. Tonight: more tired, more on your side. Numbers will be middling and that's fine. In a month: you have a working pattern — write your version first, edit for clarity only, never for taste. In a year: voice and reach compound together instead of trading off.

You don't ship. Both tabs go to sleep.

Most editing decisions made in front of metrics are rationalizations. The taste-system that's running you at 11pm after a feed-heavy day is not the same taste-system you'll have at 8am off-platform. Sleeping on a draft isn't laziness — it's getting your judgment back. Tonight: no shipping. A small itch that wants to publish anyway. Tomorrow: you'll see something you couldn't see last night. As a habit: anything you can't tell whether to publish, sleep on. Always.

A sentence to sit with

Whatever you shipped or didn't, name where the editing came from. If the algorithm did most of it, that's information. If you did, that's information. The trees the people you'll respect in five years are growing right now.

Digital Life

Is this my voice, or the algorithm's?

You have two versions of the same piece. One is what you'd write if no one was watching. The other is what you'd write if you wanted it to land. They're not the same. They used to be. The drift has been quiet.

🎙️ Creator at the edit pass 6 min

A guided self-diagnosis — no signup, nothing leaves your device.

Loading Psychology Insights...

Is this my voice, or the algorithm's?